Male and female faculty members show distinct gender bias when they assess applications from male and female students, a new study from Yale University has shown. This is an important (if depressing) study, carefully carried out and clearly reported. But how exactly did the researchers choose the names of their ‘applicants’, Jennifer and John? I’m still trying to find out.
There has been much online discussion of the new study from Yale entitled ‘Science faculty’s subtle gender biases favour male students’. The paper can be read here (open access – I’m happy to say): http://www.pnas.org/content/early/2012/09/14/1211286109.full.pdf
Updated 1 Oct 2012, see end of post.
To summarize the study: 127 academics in US science departments each analysed a student’s application for the post of lab manager (a junior position, where the applicant had the intention to apply for a doctorate in the future). The academics were asked by the researchers to assess the competence and hireability of the candidate and state what starting salary they would offer if they were to employ that candidate. They also assessed the ‘likeability’ and the degree of career mentoring they would offer candidate. What the academics were not told was that all the CVs were identical apart from the name and gender of the applicant: half were said to be from a male applicant ‘John’ and half from a female applicant ‘Jennifer’.
The results showed clear gender bias: ‘Jennifer’ was seen as less hireable, less competent, and worthy of lower salary and (particularly worrying, I feel) less mentoring than ‘John’ … although she was more ‘likeable’! It didn’t matter whether the academics who assessed the CVs were male or female, senior or junior, or from a background of chemistry, biology or physics; all showed similar bias.
Looking at this study, the methodology seems excellent, and it is almost all very clearly described.
You can read the CVs and covering letter in the supplementary materials, and they are very convincing. I had wondered how you could make identical back stories fit both male and female applicants, but in this case just excerpts are taken from the resume and references which seem to fit both genders equally.
But how were the names ‘Jennifer’ and ‘John’ chosen? Apart from the gender ‘M’ and ‘F’, the names are the only things different about the study and so the test could be affected by perceptions of those particular names. Of course, the authors have considered that issue. According to the paper, John and Jennifer are ‘two names that have been pretested as equivalent in likability and recognizeability (50).’ I have read reference 50 (Brescoll VL, Uhlmann EL (2008) Can angry women get ahead? Status conferral, gender, and workplace emotion expression. Psychol Sci 19:268–275) which is very interesting in itself (and freely available online), but which does not seem to discuss the issue of names at all. Have I missed something obvious?
The gender bias result seems very solid, but it’s annoying that I can’t find anything about how the names were ‘pre-tested’. I’ve emailed Jo Handelsman, the corresponding author, to ask and will update this post if I learn more. I expect that (if I’ve not missed something in ref 50) there is a published study which does test these names – though my searches have not found it so far.
Could the names Jennifer and John be perceived very differently? There are kings named John – could ‘John’ appear more solid and reliable than ‘Jennifer’? Would Jennifer’s Welsh form ‘Guinevere’ have given her the necessary gravitas to win the job?
But this is only a minor point about one aspect of the description of the methodology. Far more important is what this study has made me think about regarding my own unconscious biases when reading applications and writing references. The last two people I had a say in appointing were female, I’m glad to say, but that’s no grounds for complacency.
Update 1 October 2012: Lead author of the study, Corinne Moss-Racusin kindly replied to my email and said that the names John and Jennifer were pre-tested (along with many other names) by co-author Victoria Brescoll as part of another study: Brescoll and Uhlmann, 2005, Attitudes Toward Traditional and Nontraditional Parents, Psychology of Women Quarterly, 29, 436–445. The pre-testing is only briefly referred to in this paper, but I guess not everything gets written up in detail. Thanks to Corinne for her rapid and responsive reply.